[
https://issues.apache.org/jira/browse/AMBARI-22916?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siddharth Wagle updated AMBARI-22916:
-------------------------------------
Description:
We are pulling in Name Node federation support in HDFS/HDP 3.0, requires Ambari
support for NN Federation. This is a major pain point for many customers and is
part of.
Details
======
What NN federation allows is to share the Data Nodes between a set of
independent Name Nodes. Each NN has its own standby and has the meta-data
pointing to the blocks in the data nodes, however, they might share the Data
Nodes. This will let us scale to 250M files for each NN and add more NNs as we
go beyond 250M files.
Ambari support will be required to make sure we can manage and configure NN
federation.
was:
We are pulling in Name Node federation support in HDFS/HDP 3.0 urgently,
however, all our customers requires Ambari support for NN Federation. This is a
major pain point for many customers and is part of PayPal's get well plan as
well will be required for the large Daimler's deal.
Details
======
What NN federation allows is to share the Data Nodes between a set of
independent Name Nodes. Each NN has its own standby and has the meta-data
pointing to the blocks in the data nodes, however, they might share the Data
Nodes. This will let us scale to 250M files for each NN and add more NNs as we
go beyond 250M files.
Ambari support will be required to make sure we can manage and configure NN
federation.
> BE: Support Name Federation in Ambari
> -------------------------------------
>
> Key: AMBARI-22916
> URL: https://issues.apache.org/jira/browse/AMBARI-22916
> Project: Ambari
> Issue Type: Epic
> Components: ambari-server
> Affects Versions: 2.7.0
> Reporter: Paul Codding
> Assignee: Siddharth Wagle
> Priority: Critical
> Fix For: 2.7.0
>
>
> We are pulling in Name Node federation support in HDFS/HDP 3.0, requires
> Ambari support for NN Federation. This is a major pain point for many
> customers and is part of.
> Details
> ======
> What NN federation allows is to share the Data Nodes between a set of
> independent Name Nodes. Each NN has its own standby and has the meta-data
> pointing to the blocks in the data nodes, however, they might share the Data
> Nodes. This will let us scale to 250M files for each NN and add more NNs as
> we go beyond 250M files.
> Ambari support will be required to make sure we can manage and configure NN
> federation.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
